Democrats may go after data centers. That's not a reason to dump all your AI stocks
Democrats won big on Tuesday in the U.S. midterm elections, and while the party’s opposition to data centers has been spooking Wall Street, it may actually be a good thing for the artificial intelligence trade. It’s counterintuitive, but analysts and economists were arguing on Wednesday that political pushback to the data center buildout may help to productively deflate what is widely acknowledged to be a market bubble in tech resulting from AI-related investment. This could make the investment boom more sustainable in the longer term while broadening it out into other areas of the economy, even though it may add chop to the trading waters and contribute to a down-turn in the market over the short to medium term. “It is necessary that the investment in AI slow to a more sustainable pace. It is in the interest of the AI trade itself,” RSM US chief economist Joseph Brusuelas said in a Wednesday analysis. “In all bubbles there is mal-investment, malfeasance, and failures. This boom will experience its bust and a slow deflation of the bubble is clearly in the interest of a broader and more sustainable AI trade.” Markets were breathing a sigh of relief on Wednesday with all the major indices up. The S & P 500 was up 2% as of 12:00 p.m. ET; the Dow was up about 3% and the Nasdaq Composite was ahead 1.5%. Notably, the capital spenders in the AI trade were being rewarded on Wednesday as hurdles to the buildout could improve their sagging financial statements and bolster cash flows at a time when finance capital for the buildout has been growing scarce, as Nvidia CEO Jensen Huang recently observed. Hyperscalers Meta , Google , Amazon and Microsoft were all up more than 4% as of 2 p.m. AMZN 5D mountain Amazon 5 day Conversely, the buyers of that capital spend — notably power companies — are being hit hard. GE Vernova is down about 10% and Constellation Energy is down 6%. Like almost all previous market bubbles related to new technologies, over-investment of capital is almost certainly a factor in the current AI boom. Economists think the check on data center expansion from Democrats could help to let some air out of the bubble without popping it with a single prick. “A slower buildout isn’t entirely a bad thing,” Aditya Bhave, head of U.S. economics research at Bank of America, said in a Wednesday note. “The AI-related categories of capex have been growing at a 26% annualized pace in real terms since the beginning of last year. Such a front-loaded capex wave is symptomatic of over-investment. It means we face a greater risk of a hard landing down the line.” He added that “adding a bit of sand in the gears might actually benefit markets and the economy in the medium term. The jury’s still out.” Slowing the buildout could also push investment into other areas of the economy in a potentially stabilizing way. Brusuelas said it could “result in a rebalancing and redirection of capital into the real economy.” Not all sectors of the AI trade were doing well on Wednesday. The financially fragile, AI-dedicated neoclouds were getting hit hard, with CoreWeave down about 9%, Nebius down 6%, and IREN down 11%. Research firm Gartner predicted in October that 80% of neoclouds could cease independent operations by the end of 2030 due to insufficient AI revenues at the consumer and enterprise level, translating to a dearth of profitable returns for some investments. Starwood Capital’s Barry Sternlicht made a similar point on Tuesday, saying that some neoclouds could fail. Credit defaults swaps on CoreWeave – a form of risk insurance on their bonds – widened a bit on Wednesday morning, with the 5-year rising by 3.5%. Companies across the AI trade are still hungry for capital with hyperscalers, equipment providers, and algorithm companies seeking hundreds of billions over the past two weeks, according to Evercore. They want this money quickly and political roadblocks could make it harder to put it to efficient use. “The pace and scale of activity have put the cost of capital and funding capacity in focus,” Julian Emanuel at Evercore said in a Wednesday note. “While strong oversubscription suggests capital remains available, rising borrowing costs, dilution concerns, and reliance on structured financing have weighed on the space.” Even as Wall Street analysts and the big stock indices were signaling encouragement on Wednesday, the overall environment for the AI trade remains “challenging,” according to 22V Research. This is especially true in Texas, which has 27% of the total U.S. data center gigawattage, according to Wells Fargo, and whose newly elected next governor Gina Hinojosa (D) is expected to be tougher on data centers than her predecessor Greg Abbott (R). “While the macro indicator is multiple states moving towards data center pauses, the size of the buildout in Texas and the likely additional months of delays and uncertainty around the new governor will be a key focus for stocks tied to the state,” Dauvin Peterson at 22V said in a Wednesday note. “At a higher level, the AI story will move forward – and from current levels, today’s reaction and capability to compartmentalize will be telling.”
